Hi, I've been studying Benford's law lately. The problem I've been getting, however, is that the anomalies compared to the theoretical values are quite big. My last investigation was about nuclear levels between 1 MeV and -I think- 1 000 000 MeV. I got this information on this website: http://cdfe.sinp.msu.ru/services/ensdfr.html. Using this website, I managed to get ~133 000 nuclear levels, which should be enough to get rid of big statistical fluctuations. My measurement: http://desmond.imageshack.us/Himg4/scaled.php?server=4&filename=graphbenfordnuclearleve.png&res=medium [Broken] The circles show the theoretical values, the triangles show what I measured. As one can see, I really have the right trend for numbers 4 to 9. I am however worried about the fluctuations in numbers 1 to 3, they're really too big (for my intuition) to be -just- statistical fluctuations. I studied the Wikipedia page and I cannot find any argument why Benford's law would not apply on my set of data. I wonder if anyone on this forum knows why my data shows this big of anomalies vs. Benford's law? Thanks, Trinitiet. P.S. My code was written in R. I tested the code multiple times on small sets of data so I'm quite confident there's nothing wrong with my results.